A Ponceau Staining based Dot-Blot Assay for reliable and cost-effective Protein Quantification

Reliable quantification of protein extracts from tissues can be a challenge e.g. due to interference of the high fat content in tissues of the nervous system. Further problems like under- or overerstimation of protein concentrations in protein quantification kits like the bicinchoninic acid (BCA) assay can occur. In addition, common lysis buffers such as RIPA buffer are known to be unable to solubilize a large amount of proteins (~10-30%) leading to unsatisfactory and unreliable experimental results with techniques such as immunoblotting. In this work, we have developed a Ponceau S staining based protein quantification assay. This assay is compatible with tissues or cells directly lysed in 2x SDS gel loading buffer, containing bromophenolblue, leading to more complete protein extraction. Protein concentrations of several samples can be determined in a fast and cost-effective manner and subsequent experiments (e.g. Western blot) can be performed without loss of proteins. The presented protein quantification method is highly reliable, fast and economical. Using this method, it is possible to save between 2300 to 3200€ per 1000 lysates as compared to the costs of a commercial BCA kit.


204
Because the sizes of the resulting dots were unequal we used the "Integrated Density" of 205 each dot which is the product of the selected area and its mean grey value over the 206 measured area. By using the same rectangle size when analysing the dots in Fiji only the 207 grey values decrease as the protein amount increases in a linear manner (Fig. 1B, left). The

218
This is known in the literature as "coffee rings" (8). This could not be prevented by washing 219 the membrane after spotting the BSA onto it, to remove potentially interfering SDS. It has 220 been shown that the formation of such a "coffee rings" depends on the evaporation speed of 221 the liquid and the particle movement. We can only speculate about the cause of this

232
This is probably the result of loss of diluted BSA during serial dilution of the standard on the 233 walls of pipette tips and tubes (10). Therefore, we used undiluted BSA to generate standard 234 curves. For comparison, we also generated the standard curve by bicinchoninic acid (BCA) 235 assay, a well established method. The BCA assay also showed a good linearity within a 236 range of 125 ng to 2 µg. (Fig. 1B, right). Comparison of the correlation coefficients demonstrates that the PDB assay is in performance completely equal to a BCA assay ( 243 To test the applicability of our PDB method with tissue lysates, we collected spleens as 244 protein-rich organs from four different mice, lysed them in 1 ml RIPA buffer and used 1 µl per 245 dot of this lysate for quantification ( Fig. 2A).

260
In parallel, the same lysates were quantified by BCA assay for comparison. Due to the high 261 protein content in the spleen lysates, the BCA assay showed values around the upper border 262 of the standard range, between 1,7 µg/µl and 2,4 µg/µl. In contrast, with the PDB assay the 263 protein concentrations were determined between ~6 and 9 µg/ µl, three to four times higher 264 than the values given by the BCA method (Fig. 2B) (Fig. S3).

286
We hypothesized that the decreased staining effectivity might be due to the high 287 concentration (2%) of SDS in the 1:1 diluted sample which could interfere with the binding of 288 the Ponceau S dye to proteins. Therefore, we washed the membrane after drying three times 289 for five minutes in DI-tap water before staining. This led to effective staining of the dots which 290 were, compared to dots of undiluted BSA, weaker in their intensities but spread over a larger 291 area. This was probably due to the fact that two times the volume of undiluted BSA was used 292 to achieve equal protein amounts (Fig. 3A). Interestingly, we did not again observe the 293 "coffee ring"-phenomena as we did when we diluted the BSA in RIPA buffer. Again we can 294 only speculate and explain this by the reported observation that higher amounts of SDS 295 normalize the diffusion coefficient of ovalbumin which was decreased by low amounts of in Fig.1 to produce a standard curve of BSA diluted in 2x SDS LB. This mean standard curve 299 also again displayed good linearity, reflected by a mean R 2 of 0,9945 (Fig. 3B). Compared to 300 the other correlation coefficients shown in Fig. 1C, there were no changes between all three 301 different methods in linearity of prepared standards as adressed by the correlation 302 coefficients (Fig. 3C).

320
To test the suitability of direct tissue lysis in 2x SDS LB and to compare the extraction ability 321 with RIPA, we used sciatic nerves and a brain from C57/BL6 mice. These tissues are 322 normally hard to lyse due to their high content of fatty myelin. We pooled the sciatic nerves 323 from the left and right side of two different mice (left sciatic nerves were lysed in RIPA buffer 324 and right sciatic nerves were lysed in 2x SDS LB) and also used one mechanically disrupted 325 mouse brain which we divided into two halves and subsequently lysed either in RIPA buffer 326 or 2x SDS LB. Since we expected very high concentrations for the lysed brain, we also 327 included 8 µg of BSA into the range of our standard. The resulting standard curve still maintained a good linearity (Fig. S4), supporting the suitability of Ponceau S dye for the

333
With both buffers we could lyse and determine protein concentrations effectively of brain 334 pieces and nerves (Fig. 3D). To test whether quantification of protein concentrations inside 335 lysates, in which proteins were differentially extracted, gave us in both cases true values, we 336 subjected the lysates of 50, 25 and 15 ug protein content to a SDS-PAGE followed by protein 337 transfer to the membrane, subsequent Ponceau S staining and immunoblotting .

338
The Ponceau S staining showed overall equal loading between the two extraction methods if 339 one compares only sciatic nerves or brain lysates among themselves (Fig. 3E). This proves 340 similar performance of the PDB assay with either lysates prepared in in RIPA buffer or in 2x 341 SDS LB (Fig. 3E). Intriguingly, there was a general difference between the loading of sciatic 342 nerves and brain lysates (Fig. 3E). We suppose that this is due to the high abundance of 343 albumin (strong band below 70 kDa) and IgG heavy (strong band slightly above 55 kDa) and 344 light (strong band between 25 and 35 kDa) chain in the PNS which are absent in the CNS 345 due to the blood brain barrier (12, 13). The presence of these highly abundant serum 346 proteins would lead to overestimation of the real protein content of the sciatic nerve itself and 347 therefore leads to unequal loading compared to both brain lysates. This is an important point 348 if researchers attempt to compare expression of different proteins between CNS and PNS.

349
As expected, subsequent immunoblotting revealed better extraction of different proteins by 350 2x SDS LB. It has been described that e.g. cytoskeleton associated proteins and 351 extracellular matrix components are to a certain degree insoluble in RIPA buffer(11). The

352
tumor suppressor protein merlin as a cytosekeleton associated protein was extracted more in 353 2x SDS LB in both sciatic nerve and brain lysates as described before (14) (Fig.4). higher P-ERK1/2 but massively higher P-MEK1/2 signals suggesting more efficient 359 phosphatase inhibition in 2x SDS LB, probably due to the strong denaturing effect of SDS.

360
This finding is important for researchers studying fast-changing signaling processes e.g. 368 Immunoblots for indicated target proteins of the membrane shown in Fig.3E

388
Since we have shown that direct lysis of tissues in 2x SDS LB is not only compatible with our 389 method but rather even more recommended because it extracts and solubilises different 390 proteins better (Fig.4), as it has been reported previously(11), we also calculated how much 391 money could be saved with the usage of 2x SDS LB instead of RIPA buffer with addition of 392 phosphatase and protease inhibitors. One ml RIPA lysis buffer with phosphatase and 393 protease inhibitors routinely used in our laboratory costs around 2,41€ while the same 394 volume of selfmade 2x SDS LB only costs 0,1€. One ml of selfmade RIPA lysis buffer with 395 phosphatase and protease inhibitors would still cost 1,48€.. If we project this to 1000 these amounts shows that the laboratory could save 2331,67 and 3261,67€ per 1000 lysates

402
Although a similar principle was described previously (17), our study highlights some more 403 critical points and adds a new improvement: If using RIPA buffer as lysis buffer it is extremely 404 important not to dilute the BSA which is used for preparation of a standard curve. This also